PROJECT SUMMARY/ABSTRACT The goal of this proposal is to develop an objective, standardized, and efficient alternative to current methods of functional assessment for older adults across the spectrum of cognitive aging, from preclinical to mild dementia. Mild functional difficulties are a strong predictor of future cognitive decline and may precede impairment on traditional cognitive tests. Functional ability level is a criterion that distinguishes MCI from mild dementia, and unlike scores on tests of cognitive abilities, functional status generally is not as strongly influenced by race/ethnicity and socioeconomic status (SES). Consequently, information on functional status is often useful for determining the validity of low cognitive test scores in non-majority elders or elders with low educational attainment. The current benchmark standard method of functional assessment is informant report, which is subjective in nature and requires a knowledgeable partner. Thus, informant reports may be biased by cultural beliefs, lack of objectivity (e.g., they are based on individuals' personal task demands and informants' attention to detail), and in some cases may be unavailable. Traditional performance-based assessments offer advantages over informant reports, but they are time consuming to administer and score. Aim 1 of this proposal will evaluate the validity and reliability of a novel, portable, non-immersive, virtual- reality paradigm, the Virtual Kitchen Challenge (VKC). The VKC requires participants to complete two everyday tasks in a virtual kitchen using a touch-screen interface to select objects and sequence steps. A detailed automated score report reflecting time to completion, task accomplishment and performance efficiency is generated immediately after task completion. 180 individuals aged 65 and older with a range of cognitive abilities (healthy n = 60, mild cognitive impairment, MCI n = 60, mild dementia n = 60) will be administered the VKC as well as traditional cognitive tests (yielding demographically adjusted scores) and validated performance-based tasks with real objects. Informants also will be asked to complete questionnaires of functional status. We predict that the automated VKC score output will yield valid, reliable, and detailed characterizations of everyday function for participants at all levels of cognitive ability. Aim 2 will test the hypothesis that because of its practical, ecologically valid nature and objectivity, performance-based assessment of functioning (i.e., VKC) will be less influenced by race/ethnicity, education, and socioeconomic status than standard measures of cognition and informant report of function. However, if our hypothesis is not supported, demographic adjustments for interpreting VKC scores will be identified and applied. This proposal will yield an efficient and sensitive measure of everyday function for diverse older adults that may be used clinically when informants are unavailable or potentially unreliable, and may be scalable for large longitudinal studies of cognitive aging and dementia.